rapid selection against inbreeding in a wild population of a rare frog

ORIGINAL ARTICLE

Rapid selection against inbreeding in a wild populationof a rare frogGentile Francesco Ficetola,1,2 Trenton W. J. Garner,3 Jinliang Wang3 and Fiorenza De Bernardi1

1 Dipartimento di Biologia, Universita degli Studi di Milano, Milano, Italy

2 Dipartimento di Scienze dell’Ambiente e del Territorio, Universita degli Studi di Milano-Bicocca, Milano, Italy

3 Institute of Zoology, Zoological Society of London, London, UK

Introduction

Small, isolated populations are particularly subject to

inbreeding, and are thus intrinsically vulnerable to extir-

pation (Keller and Waller 2002). Both laboratory and

field studies have shown that increased inbreeding associ-

ated with small and isolated populations can strongly

reduce both the variance and mean of traits that are

important components of fitness, including survival,

fertility and individual growth rate (Saccheri et al. 1998;

Keller and Waller 2002; Rowe and Beebee 2003; Frank-

ham 2005; Wright et al. 2008). The reduced survival of

inbred individuals can further reduce population size and

increase susceptibility to stochastic processes, such that

mutual reinforcement among environmental, demo-

graphic and genetic stochasticity creates an ‘extinction

vortex’ (Gilpin and Soule 1986). However, if enough

genetic variability is retained by the population, less

inbred individuals, or individuals not carrying deleterious

alleles can be favoured through selection. If inbreeding

depression is mainly caused by the effect of strongly dele-

terious, recessive alleles, natural selection may be particu-

larly effective against them when they occur in the

homozygous state (Wang 2000). Two processes can thus

occur in small populations. First, the loss of inbred indi-

viduals can cause a further reduction of population size,

leading to cycles of exacerbated inbreeding and drift in

subsequent generations (referred to as ‘mutational melt-

down’; Lynch et al. 1995). On the other hand, selection

against inbred genotypes can purge inbreeding depression

(Hedrick 2001). In this case, populations may be able to

recover demographically despite initial demographic

losses by removing inbred individuals and reducing the

frequency of deleterious alleles. Selection can be most

effective at countering inbreeding depression under par-

ticular conditions, such as when the fitness disadvantages

Keywords

genetic purging, heterozygosity,

microsatellites, Rana latastei, temporal

variation.

Correspondence

Gentile Francesco Ficetola, Dipartimento di

Biologia, Universita degli Studi di Milano, Via

Celoria 26, 20133 Milano, Italy.

Tel.: +39 (0)2 64 48 29 45;

fax: +39 (0)2 64 48 29 96;

e-mail: [email protected]

Received: 13 April 2010

Accepted: 14 April 2010

First published online: 10 May 2010

doi:10.1111/j.1752-4571.2010.00130.x

Abstract

Populations that are small and isolated can be threatened through loss of fit-

ness due to inbreeding. Nevertheless, an increased frequency of recessive homo-

zygotes could increase the efficiency of selection against deleterious mutants,

thus reducing inbreeding depression. In wild populations, observations of evo-

lutionary changes determined by selection against inbreeding are few. We used

microsatellite DNA markers to compare the genetic features of tadpoles imme-

diately after hatch with those of metamorphosing froglets belonging to the

same cohort in a small, isolated population of the threatened frog Rana latastei.

Within a generation, the inbreeding coefficient (FIS) decreased: at hatch, FIS

was significantly >0, whereas FIS was <0 after metamorphosis. Furthermore,

heterozygosity increased and allelic frequencies changed over time, resulting in

the loss of genotypes at metamorphosis that were present in hatchlings. One

microsatellite locus exhibited atypically large FST values, suggesting it might be

linked to a locus under selection. These results support the hypothesis that

strong selection against the most inbred genotypes occurred among early life-

history stages in our population. Selective forces can promote changes that can

affect population dynamics and should be considered in conservation planning.

Evolutionary Applications ISSN 1752-4571

30 ª 2010 Blackwell Publishing Ltd 4 (2011) 30–38

of deleterious mutations are particularly strong, inbreed-

ing does not occur randomly, and migrants are not

(re)introducing deleterious alleles (Glemin 2003; Edmands

2007).

Detecting the ongoing population genetic processes that

lead to mutational meltdown or recovery can provide key

insights for the management of threatened species at greater

risk of inbreeding. However, detecting either of these

patterns in wild animal populations can be challenging.

Under natural settings, measuring the relative performance

of individuals with respect to their inbreeding status usually

requires long-term capture/mark/recapture analyses over

multiple generations and the reconstruction of pedigrees.

For this and other reasons, the measurement of inbreeding

depression in the wild has been done primarily in larger

vertebrates where parentage is easier to determine and the

number of offspring per female is small (Keller and Waller

2002; Kruuk and Clutton-Brock 2008). For taxa where such

parameters are not so amenable to measurement, two other

approaches are used. Inbreeding depression can be

measured under laboratory conditions, however inbreeding

depression is frequently stronger under natural (stressful)

conditions, and results obtained by laboratory studies may

not reflect the complexity of natural conditions (Bijlsma

et al. 1999; Joron and Brakefield 2003; Halverson et al.

2006). Alternatively, averaged population responses (e.g.,

FIS) can be used, by comparing populations with different

inbreeding histories rather than measuring fitness of

individuals (Saccheri et al. 1998).

Molecular genetics can be a powerful approach for the

study of microevolutionary changes over a range of tem-

poral scales, from a few generations to thousands years.

In anuran amphibians, typically 80–95% of mortality

occurs between hatch and metamorphosis (Vonesh and

De la Cruz 2002) and inbreeding depression may be most

readily detected through the measurement of relevant

population genetics parameters during these periods of

heavy mortality (Rowe and Beebee 2003; Halverson et al.

2006; Chapman et al. 2009). In this study, we used

microsatellite DNA polymorphisms to study temporal

variation of genetic features in small, isolated population

of a threatened frog suffering strong inbreeding depres-

sion (Ficetola et al. 2007). We analyzed the genetic fea-

tures of the same cohort at two subsequent life-history

stages, immediately after hatch and at metamorphosis;

our sampling protocol was therefore designed to compare

the genetic features of a cohort before and after the effects

of potentially strong natural selection. We hypothesized

that strong selection against the most inbred genotypes

might cause strong shifts in genetic features of the popu-

lation. Evolutionary signatures of selection against

inbreeding include an increase in heterozygosity, a

decrease in the inbreeding coefficient FIS, and loci exhibit-

ing outlying values of FST (Beaumont 2005). Moreover,

the reduced frequency or even elimination of specific

alleles all may occur if selection is removing deleterious

alleles expressed in the homozygous condition.

Methods

Study species and area

The Italian agile frog, Rana latastei Boulenger, is endemic

to Northern Italy and adjacent countries and is classified

as vulnerable to extinction by the IUCN (Andreone et al.

2008). Habitat requirements are river washes, forest

brooks and ponds associated with lowland forest cover,

primarily along the Po River drainage in Italy. In this

area, habitat has been subject to intense urbanization and

agriculture. In particular, the western part of the distribu-

tion of R. latastei is broken into a limited number of

small and isolated forest fragments. Many breeding sites

are characterized by small census size and geographic iso-

lation (Ficetola et al. 2007) and this combination of isola-

tion and small population size has increased the risk of

extinction of populations (Ficetola and De Bernardi

2004). Further, the part of the species range from which

our study population comes from is characterized by a 3·reduction in genetic variability when compared to the

eastern part of the range (Garner et al. 2004) and experi-

ments have shown that reductions in genetic variability

correlate well with reduced fitness (Pearman and Garner

2005; Ficetola et al. 2007). Breeding occurs in late winter;

females lay a single egg mass per year (Barbieri and

Mazzotti 2006), therefore the number of clutches repre-

sents both annual population reproductive output and

the number of breeding females in a population. Multiple

paternity has never been observed in R. latastei. Some

studies have described multiple paternity in related spe-

cies, but the offspring of polyandrous mating is a small

proportion. For instance, in Rana temporaria, clutch

piracy affected 84% of clutches, but pirate males fertilized

only 26% of pirated clutches, and in these clutches 24%

of embryos were fertilized by pirates, i.e., 5% of all eggs

were sired by a secondary male (Vieites et al. 2004). Simi-

larly, in Rana dalmatina only 4% of all eggs are sired by

a secondary male (Lode et al. 2005). Our results are

robust even with the removal of 5% of larvae from the

analyses, therefore the occurrence of multiple paternity at

rates similar to the ones observed in related frogs would

have minimal influence on the outcome of this study

(see also Ficetola and De Bernardi 2009; Ficetola et al.

2010 for further details).

We studied one population located close to the Alserio

Lake, in Northern Italy (population AL in the map in

Ficetola et al. 2007). At this location, we detected a total

of 43 clutches during the 2003 breeding season, most of

Ficetola et al. Rapid selection against inbreeding

ª 2010 Blackwell Publishing Ltd 4 (2011) 30–38 31

which were found in a single pond (hereafter: study

pond; 32 egg masses in 2003). The study pond is small

(surface area 60 m2), shallow (maximum water depth:

50 cm) and with clear water; clutches are deposited onto

submerged tree roots and sticks below the water surface,

thus, they are easily detected. Clutch counts therefore

provide a reliable standard census method for agile frogs

(Salvidio 2009). A few clutches were found in a small

satellite wetland within 50 m of the study pond; clutches

from the satellite pond were not considered here. The

study population is isolated: the next available breeding

location was located 3 km away, and the intervening

distance was composed of unsuitable habitat, lacking

both suitable breeding locations and contiguous forested

habitat; a motorway acts as a further barrier.

Sampling protocol and analyses

We subsampled all 32 viable clutches deposited in the

study pond in March 2003, taking a single tadpole per

egg mass for DNA extraction. This sample therefore rep-

resented the entire hatching cohort. In late June–early

July, we returned to the study pond and we sampled the

pond banks to collect metamorphic froglets (stage 39–42;

Gosner 1960). We captured active individuals; we also

used a net to collect leaf litter and cryptic individuals;

overall, we obtained 29 froglets. We collected a small tis-

sue sample (remnant tail tip) from each metamorph for

DNA extraction. We restricted our sampling to these

semiaquatic stages to prevent sampling any possible meta-

morphs that may have come from the adjacent woodland.

Overall, 78 microsatellite loci developed either for

R. latastei or for other Rana species have been tested on

R. latastei samples. Thirty-one of these loci were mono-

morphic, and 41 did not amplify or showed presence of

null alleles, therefore only six polymorphic loci are actu-

ally available for this species, a proportion significantly

lower than in other frog species (Garner and Tomio

2001; Primmer and Merila 2002; T.W.J.G. unpublished

data); this is probably caused by the colonization history

of this small range species combined with the extreme

landscape fragmentation (Garner et al. 2004; Ficetola

et al. 2007). DNA was extracted using a Chelex-based

protocol and amplified at variable microsatellites follow-

ing Garner and Tomio (2001). [33P]-labeled fragments

were electrophoresed through 6% polyacrylamide gels and

fragments were detected using autoradiography. Alleles

were visually scored against a pGem�-3zf+ (Promega,

Madison, Wisconsin, USA) sequence reference marker.

Further details on genetic analyses are provided in

Ficetola et al. (2007).

We used Genepop 3.4 (Raymond and Rousset 1995) to

test for linkage disequilibrium among the microsatellite

loci that proved to be variable, and to test for Hardy–

Weinberg equilibrium in larvae and metamorphs. To

compare the genetic variability of hatchlings and of meta-

morphs, we estimated the percentage of polymorphic loci,

allelic richness, expected and observed heterozygosity for

both developmental stages. We also tested for genetic dif-

ferences among developmental stages by calculating

respective FST values, and permuting these values 10 000

times to test for significance.

We calculated FIS for hatchlings and for metamorphs

as a measure of the level of inbreeding in a cohort

(Wright 1969); we estimated 95% confidence intervals

for each coefficient by bootstrapping 10 000 times, and

tested for significant differences in the inbreeding coef-

ficient among developmental stages. We estimated FIS

separately for hatchlings and metamorphs, therefore dif-

ferences in FIS might be in principle due to differences

in the reference population considered for FIS estima-

tion. To address this, we used simulations to evaluate

whether random mortality or sampling, and changes in

the reference population, may have caused the observed

changes in FIS. We simulated 1000 individuals with the

same FIS of hatchlings and the same allele frequencies

detected in the hatchlings. Subsequently, we randomly

selected 29 individuals from the simulated population

to form the metamorphic sample, and we estimated FIS

for these individuals. This process was repeated

1 000 000 times.

We used Fdist2 (Beaumont and Nichols 1996; Beau-

mont 2000) to identify loci showing the potential signa-

ture of selection. This approach is based on the principle

that loci under selection or tightly linked to a locus under

selection should exhibit outlying FST values (outliers)

(Beaumont 2005). We generated a null distribution based

on 30 000 simulated loci; the mean FST was calculated on

the basis of empirical distribution and then iteratively

modified as outliers were identified and removed (Beau-

mont and Nichols 1996). To evaluate the robustness of

our results to different approaches in the estimation of

expected FST, this analysis was repeated by calculating FST

without the iterative removal of outliers, which yields

more conservative results. This second analysis produced

very similar results (not shown). Results from Fdist2

may be misleading when analyzing multiple populations

grouped in phylogenetic hierarchies (Latta 2008). For

analysis with Fdist2, we treated the larval cohort as the

first ‘population’, and the metamorphic cohort as the sec-

ond ‘population’. Therefore, phylogenetic hierarchies are

absent from our data.

Finally, we used a Correspondence Analysis (CA) to

detect differences in genotype distributions between tad-

poles and metamorphs by graphically projecting all the

individuals on the space defined by factors extracted on

Rapid selection against inbreeding Ficetola et al.


the basis of the similarity of their allelic states. This last

analysis was completed using Genetix 4.05 (Belkir 2004).

Results

Three loci (RlatCa17, RlatCa21 and Rt2Ca9) were mono-

morphic across the entire sample: therefore all subsequent

analyses were performed using the remaining three poly-

morphic loci (RlatCa9, RlatCa27 and RlatCa41). We did

not detect evidence of linkage disequilibrium among any

of these loci (all P > 0.08). In the earlier (larval) life-

history stage, analysis of Hardy–Weinberg equilibrium

showed a nonsignificant trend for heterozygote deficiency

(P = 0.07). When loci were examined individually, only

locus RlatCa27 exhibited a significant heterozygote deficit

in larvae (P = 0.004). Conversely, in the later, postmeta-

morphic life-history stage we observed a significant global

heterozygote excess (P = 0.013), despite the lack of any

effect when any one locus was analyzed independently (all

P > 0.1). Percentage of polymorphic loci, allelic richness

and expected heterozygosity were similar in both stages.

Observed heterozygosity of the cohort increased from

0.404 at hatch to 0.547 in metamorphs; nevertheless, stan-

dard errors were large (Table 1). At hatch, the inbreeding

coefficient was significantly >0 (FIS = 0.183, 95% CI:

0.027 to 0.305). In contrast, the inbreeding coefficient of

metamorphic individuals was negative (FIS = )0.169, 95%

CI: )0.027 to )0.337, Fig. 1) and significantly different

from the hatch value (P < 0.001). This decrease in the

inbreeding coefficient over life-history stages was strongly

influenced by one locus (RlatCa27) but was evident at all

three of the variable loci (Fig. 1). Exclusion of locus

RlatCa27 from the analysis resulted in a marginally non-

significant trend (P = 0.057). Simulations showed that the

observed changes in FIS could not be explained by ran-

dom mortality/sampling, as the observed FIS of metamor-

phs was lower than the first percentile of the distribution

of simulated data (Fig. 2).

Genotype frequencies differed significantly among life-

history stages (FST = 0.101, P < 0.001), but pairwise FST

was not significant if locus RlatCa27 was removed from

the analysis (FST = 0.010, P = 0.169). Locus RlatCa27

exhibited an atypically large FST value among life-history

stages (P = 0.001, Fig. 3). This result is particularly strik-

ing because Fdist2 is expected to have relatively low

power for detecting outliers when the number of loci

included in the analysis is not large (Beaumont and

Nichols 1996; Nielsen et al. 2006).

Both the significant decrease in the inbreeding coeffi-

cient and change in genotype distribution detected among

life-history stages was primarily driven by the disappear-

ance of certain homozygotes in the metamorphic sample

Table 1. Genetic features of tadpoles immediately after hatch and of

metamorphosing froglets in a population of Rana latastei. Data

derived from polymorphisms at microsatellite loci. Standard errors are

in parentheses.

Hatchlings Metamorphs

N polymorphic loci 3 3

Number of alleles 10 10

Allelic richness 3.31 3.33

Expected heterozygosity

All loci 0.484 (0.108) 0.461 (0.099)

RlatCa9 0.405 0.345

RlatCa27 0.608 0.536

RlatCa41 0.441 0.498

Observed heterozygosity

All loci 0.404 (0.075) 0.547 (0.147)

RlatCa9 0.423 0.379

RlatCa27 0.321 0.607

RlatCa41 0.469 0.655

–1

–0.5

0

0.5

1

Loci

FIS

Ca9 Ca27 Ca41 All loci

Figure 1 Inbreeding coefficient (FIS) of hatchling (empty diamonds)

and metamorphic Rana latastei (solid diamonds) derived from variation

at three polymorphic microsatellite loci (RlatCa9, RlatCa27 and

RlatCa41). Error bars are 95% CI.

0–0.3 –0.2 –0.1 0.1 0.2 0.3 0.4

0.01

0.02

0.03

0.04

0.05

0.06

Simulated valuesFIS

Fre

quen

cy Observedof metamorphs

FIS

Figure 2 Distribution of simulated FIS values. The two bold vertical

lines indicate the 1% percentiles; the arrow indicates the FIS value

observed in metamorphs. See text for more details on the simulations.



that were present in the hatch data set. Specifically, 38%

of hatchlings were homozygous for the same allele

(154 bp) at locus RlatCa27, while none of the metamor-

phosing froglets were homozygous for this allele at this

locus (Fig. 4). The frequency of this allele dropped from

0.518 in larvae to 0.054 in the metamorphs. One further

homozygote was eliminated from the genotype pool as

development progressed (166 bp) and no new homozyg-

otes occurred in the metamorphic data pool that were

not previously detected in the earlier life-history stage

(Fig. 4).

The first component extracted from the CA explained

24.3% of the total variation of genotypes and showed

clearly how the genotypes of metamorphs constituted a

small subset of the hatchling genotypes (Fig. 5). Hatch-

ling scores along the first CA component ranged from

+2 to )1, whereas all metamorph scores were <0.5.

Genotype variability along the first component signifi-

cantly decreased from hatch to metamorphosis (Levene’s

test for homogeneity of variances: F1,59 = 29.0, P < 0.001,

Fig. 5).

Discussion

Our study showed that within a cohort of R. latastei FIS

decreases over time while observed heterozygosity

increased slightly. One locus was predominantly responsi-

ble for atypically large differences among life-history

stages (FST) that are considered to be evidence of selec-

tion at a locus (Beaumont and Nichols 1996). Further,

only a subset of the genotypes present among clutches

persisted through to metamorphosis and the frequency of

a common allele, present only in hatchlings in the homo-

zygous state, was strongly reduced in the metamorphic

cohort. We present arguments below for why we believe

these observed patterns are indicative of strong selection

against inbreeding.

Microsatellites are usually considered to be neutral loci.

Inferring selective processes from changes in neutral

markers requires the elimination of the alternative expla-

nations, such as immigration, drift, nonrandom mating

and mutation. In our study, nonrandom mating and

mutation can be ignored, as we sampled a single, nonre-

productive cohort. Immigration had no effect on our

sample because of pond isolation and the fact we sampled

before metamorphs from nearby wetlands could have

emigrated. Stochasticity may have a role in some of the

changes observed; for instance, the frequency of some but

not all the homozygotes decreased in metamorphs

(Fig. 4). However, stochasticity cannot be the sole cause

for the observed changes, as simulations showed that it is

highly unlikely that the strong decrease of FIS we have

observed could be attributed to stochastic processes

(Fig. 2). Thus, we conclude that natural selection was the

most likely driver of the temporal variation in genetic fre-

quencies we observed; the strong decrease of FIS suggests

0

0.25

0.5

Freq

uenc

y

0

0.25

0.5

154–154 162–162 166–166 144–162 144–166 154–162 154–166 162–166

Genotypes

Freq

uenc

y

Figure 4 Observed frequencies of genotypes at locus RlatCa27 for

the two life-history stages. Alleles are enumerated by length in base

pairs. Open bars: hatchlings; solid bars: metamorphs.

–0.05

0

0.05

0.1

0.15

0.2

0.25

0 0.25 0.5 0.75 1

FS

T

RlatCa9

RlatCa41

RlatCa27

Heterozygosity

Figure 3 Individuation of loci under selection using FDIST2: plot of FST

values against heterozygosity estimates (Beaumont and Nichols 1996)

for the comparison tadpoles/metamorphs. Each diamond represents a

microsatellite marker; the intermediate line is the expected median;

upper and lower lines are the limits of the 99% envelope.

CA

Hatchlings Metamorphs

2

1

0

–1

CA

Sco

res

(1st

com

pone

nt)

Figure 5 Box-plots representing the distribution of genotypes of

hatchlings and metamorphs along the first component extracted by

Correspondence Analysis.



that selection against inbreeding is the most likely expla-

nation of the observed genetic changes. Nevertheless, it

should be remarked only locus RlatCa27 was not in

Hardy–Weinberg equilibrium in larvae; other possibilities

therefore exist for the pattern at that locus. For example,

allele 154 might be associated with mate choice in adults,

causing an excess of homozygotes in eggs, then selected

against in larvae. We are not the first to observe a

relationship between measures of genetic diversity at

microsatellite loci and measures of fitness (Chapman

et al. 2009). Furthermore, previous research has identified

specific patterns in the correlation between marker vari-

ability and fitness that are matched by our data set. For

example, in red deer, reed warblers, harbor seals and sea

lions, heterozygosity at multilocus microsatellite geno-

types was strongly correlated with components of fitness,

but in all these studies single microsatellite loci were

shown to have a significantly greater contribution to in-

terindividual variability and locus-specific correlations

with fitness components (Pemberton et al. 1999; Hansson

et al. 2004; Acevedo-Whitehouse et al. 2006).

Balloux et al. (2004) showed that, in the absence of

strong linkage disequilibrium and complex mating sys-

tems, it should be necessary to sample large numbers of

microsatellite loci to detect a true fitness/heterozygosity

correlation. This has not prevented the relationship

between diversity estimated at microsatellites and fitness

from being detected frequently in studies using a handful

of markers. In our study species, six loci were sufficient

for detecting significant variation in genetic diversity

across the species range (Garner et al. 2004) and this vari-

ation in diversity correlated strongly with fitness measures

(Pearman and Garner 2005; Ficetola et al. 2007). The fact

that we detected an effect with only half that many loci

reflects the pattern of genetic variability inherent to the

western part of the species range. Fixation for three of six

published loci used in this study was previously detected

in the Lombardy Region because of the joint effect of

postglacial colonization and human-driven habitat

fragmentation (Garner et al. 2004; Ficetola et al. 2007);

during the course of PCR primer development and cross-

amplification with primers developed for other Rana

species, another 31 primer sets were characterized as

monomorphic in R. latastei (Garner and Tomio 2001;

Primmer and Merila 2002). Thus, based on a random

sample of dinucleotide repeat regions that amplify reli-

ably, the true number of polymorphic loci may be as low

as 3 of 37.

FIS of larvae was significantly higher than zero (Fig. 1),

which at first glances is puzzling, as FIS is a measure of

inbreeding as nonrandom mating (Keller and Waller

2002), and we considered a single population breeding in

one small pond, where it is difficult to invoke substruc-

turing (Wahlund effect). Nevertheless, large FIS values are

often observed in small, isolated populations of anuran

amphibians (e.g., Hitchings and Beebee 1997; Andersen

et al. 2004). Increased levels of polygyny are commonly

reported for anurans, and have also been described for

the study population (Ficetola et al. 2010); polygyny can

result in inbreeding coefficients greater than expected

under random mating; this effect can be exacerbated in

small populations (Balloux et al. 2004). No matter the

process originating positive FIS in larvae, the change in

FIS was more striking than its absolute values: FIS strongly

decreased during larval development, indicating than the

less inbred genotypes were favoured by selection.

Linkage between microsatellites and loci under selec-

tion can be particularly strong in small, inbred popula-

tions carrying a heavy genetic load and after bottlenecks,

as in these populations larger portions of the genome are

expected to be under balancing selection, and probability

is greatest for loci linked to overdominant loci to remain

polymorphic after bottleneck events (Ohta 1982; Balloux

et al. 2004; Van Oosterhout et al. 2004; Zartman et al.

2006). The few microsatellites retaining heterozygosity

can therefore constitute a nonrandom sample of the gen-

ome (Santucci et al. 2007). The study population corre-

sponds well to this situation: it is small, isolated and

passed through a demographic bottleneck some years

before the season in which we sampled for this study

(Ficetola and De Bernardi 2005). From this, the relative

importance of three loci in our species for describing

genetic variability is likely greater than in other, more

genetically variable species. A limited number of variable

microsatellites is typical of endangered species suffering

loss of genetic diversity (Gebremedhin et al. 2009a). These

species are often the ones for which studies are more

urgent; lack of variable microsatellites should not hamper

performing analyses that can provide important manage-

ment information (Richards-Zawacki 2009).

There is debate if the relationship between inbreeding

and fitness is due to the effects of many loci of small effect

or the effects of a few loci, or even one, of marked effect

(Chapman et al. 2009). The limited panel of available

markers available for this study does not allow us to distin-

guish between a multilocus or single locus model. Even so,

although all our polymorphic loci contribute to the

observed changes in FIS and heterozygosity, the strong

effect of locus RlatCa27 suggests that certain allele(s) at this

locus may be linked to lethal or semi-lethal allele(s) at a

locus involved in early development. The observation that

not all loci decreased their frequency in the homozygote

state (e.g., 162 bp: Fig. 4) supports the hypothesis that not

all of the microsatellite allele(s) is/are associated with dele-

terious mutations. Such linkage patterns between microsat-

ellites and genes exposed to strong selection have been



inferred and even detected several times (Hansson et al.

2004; Acevedo-Whitehouse et al. 2006; Nielsen et al. 2006;

Makinen et al. 2008; Amos and Acevedo-Whitehouse

2009), and might be more detectable in small, isolated

populations (Balloux et al. 2004; Van Oosterhout et al.

2004).

An increased frequency of recessive homozygotes

through inbreeding is expected to increase the efficiency

of natural selection against deleterious mutants and

reduce inbreeding depression through purging (Crow

1970; Boakes and Wang 2005). If inbreeding depression is

mainly caused by deleterious mutations having a strong

effect, selection may rapidly eliminate deleterious alleles,

and population fitness rebounds (Wang et al. 1999; Wang

2000) while the inbreeding levels of populations are

reduced (Moore 2007). However, whether purging

actually occurs to an extent that enables the recovery of

populations has been strongly debated, primarily due to

the equivocal nature of both field and experimental

evidence (Leberg and Firmin 2008; Chapman et al. 2009).

The disappearance of one common allele occurring pri-

marily in the homozygous state among life-history stages

(154 bp at locus RlatCa27) fits the signature of strong

selection against inbreeding. Theoretical models suggest

that this type of selection can be particularly effective in

species with large reproductive ability (Wang 2000). In

many anuran species, including R. latastei, fecundity is on

the order of thousands of eggs. Moreover, larval mortality

is high and density dependent, creating ideal conditions

for selection to take effect early during development. If a

sufficient number of females reproduce, the death of a

large proportion of tadpoles may not threaten population

persistence in the short term (Vonesh and De la Cruz

2002). This suggests that purging could be more effective

in anuran populations than in taxa with limited individ-

ual reproductive effort, such as larger vertebrates (Jamie-

son et al. 2006; Chapman et al. 2009).

Under these conditions, selection may affect population

dynamics by reducing inbreeding depression and the risk

of an extinction vortex. Relationships between evolutionary

and population dynamics can have important conse-

quences for management and in population viability analy-

ses, and when alternative options (e.g., translocations

versus in situ management) are compared (Ficetola and De

Bernardi 2005; Ficetola et al. 2007). For instance, actions

allowing for a rapid demographic rebound (Wang et al.

1999), such as habitat reclamation, can be particularly

effective for the recovery of populations. Our knowledge of

the dynamics of inbreeding depression in the wild is far

from complete, however, and more studies are required to

improve decision making that account for the genetic

responses that may or may not be associated with popula-

tion recovery. Long-term studies of amphibian populations

are increasingly available, and will be necessary to evaluate

the relative importance of selection and inbreeding in

determining population dynamics, and whether selection is

strong enough to rescue isolated populations from

inbreeding depression (Bensch et al. 2006). Our analysis

considered only one population. Populations often experi-

ence different histories, conditions and selective forces;

analyses including a larger number of populations are bet-

ter for describing general patterns (Halverson et al. 2006).

To date, many studies on temporal variation of genetic

features in wild, nonmodel species have been performed

over long temporal scales. Our analysis showed a striking

change in genetic features during a very short time, which

might affect population dynamics. The fate of the popula-

tion will be therefore determined not only by the interac-

tion between environment, demographic processes and

inbreeding depression. Selective forces can act over the

functional variability among individuals, promoting

changes influencing the future of endangered species, and

should be considered in the analysis of population dynam-

ics (Bensch et al. 2006; Saccheri and Hanski 2006). The

short time scale over which these forces can operate may

increase the complexity of interactions between environ-

ment, demography and genetics; molecular genetics can

help to elucidate these processes in the wild, also at short

temporal scales. The panel of tools available for conserva-

tion genetics is quickly broadening. The possibility to

detect the signature of selection in natural populations

using both neutral and non-neutral markers may increase

our understanding of evolutionary processes in threatened

populations (Gebremedhin et al. 2009b), providing a wid-

ening range of management options for conservation.

Acknowledgements

We thank R. Martinelli, L. Trotta and R. Pennati for

helping during laboratory work, and P. Taberlet and three

anonymous referees for comments on previous draft of

the manuscript. G.F.F. was partially funded by a scholar-

ship of the University of Milano-Bicocca.

Literature cited

Acevedo-Whitehouse, K., T. R. Spraker, E. Lyons, S. R. Melin, F.

Gulland, R. L. Delong, and W. Amos. 2006. Contrasting effects

of heterozigosity on survival and hookworm resistence in

California sea lion pups. Molecular Ecology 15:1973–

1982.

Amos, W., and K. Acevedo-Whitehouse. 2009. A new test for geno-

type-fitness associations reveals a single microsatellite allele that

strongly predicts the nature of tuberculosis infections in wild boar.

Molecular Ecology Resources 9:1102–1111.

Andersen, L. W., K. Fog, and C. Damgaard. 2004. Habitat fragmenta-

tion causes bottlenecks and inbreeding in the European tree frog



(Hyla arborea). Proceedings of the Royal Society B: Biological

Sciences 271:1293–1302.

Andreone, F., T. W. J. Garner, B. R. Schmidt, C. Corti, and M. Vogrin.

2008. Rana latastei. In S. N. Stuart, M. Hoffman, J. S. Chanson, N.

A. Cox, R. J. Berridge, P. Ramani, and B. E. Young, eds. Threatened

Amphibians of the World, 506 pp. Lynx Editions, IUCN and

Conservation International, Barcelona.

Balloux, F., W. Amos, and T. Coulson. 2004. Does heterozygosity esti-

mate inbreeding in real populations? Molecular Ecology 13:3021–

3031.

Barbieri, F., and S. Mazzotti. 2006. Rana latastei Boulenger, 1879. In R.

Sindaco, G. Doria, E. Razzetti, and F. Bernini, eds. Atlas of Italian

Amphibians and Reptiles, pp. 362–367. Polistampa, Firenze.

Beaumont, M. A. 2000. Fdist2 – a program to detect loci that might

be under selection in samples from structured populations. http://

www.rubic.rdg.ac.uk/~mab/software.html. (Accessed April 2010)

Beaumont, M. A. 2005. Adaptation and speciation: what can FST tell

us? Trends in Ecology & Evolution 20:435–440.

Beaumont, M. A., and R. A. Nichols. 1996. Evaluating loci for use in

the genetic analysis of population structure. Proceedings of the

Royal Society B: Biological Sciences 263:1619–1626.

Belkir, K. 2004. GENETIX, Logiciel Sous WindowsTM Pour la Gene-

tique des Populations. Universite de Montpellier II, Monpellier.

Bensch, S., H. Andren, B. Hansson, H. C. Pedersen, H. Sand, D.

Sejberg, P. Wabakken et al. 2006. Selection for heterozygosity gives

hope to a wild population of inbred wolves. PLoS ONE 1:e72.

Bijlsma, R., J. Bundgaard, and W. F. Van Putten. 1999. Environmental

dependence of inbreeding depression and purging in Drosophila

melanogaster. Journal of Evolutionary Biology 12:1125–1137.

Boakes, E., and J. Wang. 2005. A simulation study on detecting purg-

ing of inbreeding depression in captive populations. Genetical

Research 86:139–148.

Chapman, J. R., S. Nakagawa, D. W. Coltman, J. Slate, and B. C.

Sheldon. 2009. A quantitative review of heterozygosity-fitness

correlations in animal populations. Molecular Ecology 18:2746–

2765.

Crow, J. F. 1970. Genetic loads and the cost of natural selection. In K.

J. Kojima, ed. Mathematical Topics in Population Genetics, pp.

128–177. Springer, Berlin.

Edmands, S. 2007. Between a rock and a hard place: evaluating the

relative risks of inbreeding and outbreeding for conservation and

management. Molecular Ecology 16:463–475.

Ficetola, G. F., and F. De Bernardi. 2004. Amphibians in an human-

dominated landscape: the community structure is related to habitat

features and isolation. Biological Conservation 119:219–230.

Ficetola, G. F., and F. De Bernardi. 2005. Supplementation or in situ

conservation? Evidence of local adaptation in the Italian agile frog

Rana latastei and consequences for the management of populations.

Animal Conservation 8:33–40.

Ficetola, G. F., and F. De Bernardi. 2009. Offspring size and survival in

the frog Rana latastei: from among-population to within-clutch var-

iation. Biological Journal of the Linnean Society 97:845–853.

Ficetola, G. F., T. W. J. Garner, and F. De Bernardi. 2007. Genetic

diversity, but not hatching success, is jointly affected by post glacial

colonization and isolation in the threatened frog, Rana latastei.

Molecular Ecology 16:1787–1797.

Ficetola, G. F., E. Padoa-Schioppa, J. Wang, and T. W. J. Garner.

2010. Polygyny, census and effective population size in the threa-

tened frog, Rana latastei. Animal Conservation doi:10.1111/j.1469-

1795.2009.00306.x.

Frankham, R. 2005. Stress and adaptation in conservation genetics.

Journal of Evolutionary Biology 18:750–755.

Garner, T. W. J., and G. Tomio. 2001. Microsatellites for use in studies

of the Italian Agile Frog, Rana latastei (Boulenger). Conservation

Genetics 2:77–80.

Garner, T. W. J., S. Angelone, and P. B. Pearman. 2004. Genetic diver-

sity across a vertebrate species’ range: a test of the central-peripheral

hypothesis. Molecular Ecology 13:1047–1053.

Gebremedhin, B., G. F. Ficetola, S. Naderi, H.-R. Rezaei, C. Maudet,

D. Rioux, G. Luikart et al. 2009a. Combining genetic and ecological

data to assess the conservation status of the endangered Ethiopian

Walia ibex. Animal Conservation 12:89–100.

Gebremedhin, B., G. F. Ficetola, S. Naderi, H.-R. Rezaei, C. Maudet,

D. Rioux, G. Luikart et al. 2009b. Frontiers in identifying conser-

vation units: from neutral markers to adaptive genetic variation.

Animal Conservation 12:107–109.

Gilpin, M. E., and M. E. Soule. 1986. Minimum viable populations:

processes of extinction. In M. E. Soule, ed. Conservation Biology:

The Science of Scarcity and Diversity, pp. 19–34. Sinauer Associates,

Sunderland, MA.

Glemin, S. 2003. How are deleterious mutations purged? Drift versus

nonrandom mating. Evolution 27:2678–2687.

Gosner, K. L. 1960. A simplified table for staging Anuran embryos and

larvae with notes on identification. Herpetologica 16:183–190.

Halverson, M. A., D. K. Skelly, and A. Caccone. 2006. Inbreeding

linked to amphibian survival in the wild but not in the laboratory.

Journal of Heredity 97:499–507.

Hansson, B., E. Westerdahl, D. Hasselquist, M. Akesson, and S.

Bensch. 2004. Does linkage disequilibrium generate heterozigosity-

fitness correlations in great reed warblers? Evolution 58:870–879.

Hedrick, P. W. 2001. Conservation genetics: where are we now? Trends

in Ecology & Evolution 16:629–636.

Hitchings, S. P., and T. J. C. Beebee. 1997. Genetic substructuring as a

result of barriers to gene flow in urban Rana temporaria (common

frog) populations: implications for biodiversity conservation.

Heredity 79:117–127.

Jamieson, I. G., G. P. Wallis, and J. V. Briskie. 2006. Inbreeding and

endangered species management: is New Zealand out of step with

the rest of the world? Conservation Biology 20:38–47.

Joron, M., and P. M. Brakefield. 2003. Captivity masks inbreeding

effects on male mating success in butterflies. Nature 424:191–194.

Keller, L. F., and D. M. Waller. 2002. Inbreeding effects in wild popu-

lations. Trends in Ecology & Evolution 17:230–241.

Kruuk, L. E. B., and T. H. Clutton-Brock. 2008. Introduction. Evolu-

tionary dynamics of wild populations: the use of long-term pedigree

data. Proceedings of the Royal Society B: Biological Sciences

275:593–596.

Latta, R. G. 2008. Conservation genetics as applied evolution: from

genetic pattern to evolutionary process. Evolutionary Applications

1:84–94.

Leberg, P. L., and B. D. Firmin. 2008. Role of inbreeding depression

and purging in captive breeding and restoration programmes.

Molecular Ecology 17:334–343.

Lode, T., M.-J. Holveck, and D. Lesbarreres. 2005. Asynchronous arri-

val pattern, operational sex ratio and occurrence of multiple paterni-

ties in a territorial breeding anuran, Rana dalmatina. Biological

Journal of the Linnean Society 86:191–200.

Lynch, M., J. Conery, and R. Burger. 1995. Mutation accumulation

and the extinction of small populations. The American Naturalist

146:489–518.



Makinen, H. S., T. Shikano, J. M. Cano, and J. Merila. 2008. Hitchhik-

ing mapping reveals a candidate genomic region for natural selec-

tion in three-spined stickleback chromosome VIII. Genetics

178:453–465.

Moore, J. 2007. Phenotype matching and inbreeding avoidance in

African elephants. Molecular Ecology 16:4421–4423.

Nielsen, E. E., M. M. Hansen, and D. Meldrup. 2006. Evidence of

microsatellite hitch-hiking selection in Atlantic cod (Gadus morhua

L.): implications for inferring population structure in nonmodel

organisms. Molecular Ecology 15:3219–3229.

Ohta, T. 1982. Linkage disequilibrium due to random genetic drift in

finite subdivided populations. Proceedings of the National Academy

of Sciences of the United States of America 79:1940–1944.

Pearman, P. B., and T. W. J. Garner. 2005. Susceptibility of Italian

Agile Frog populations to an emerging Ranavirus parallels popula-

tion genetic diversity. Ecology Letters 8:401–408.

Pemberton, J. M., D. W. Coltman, T. N. Coulson, and J. Slate. 1999.

Using microsatellites to measure the fitness consequences of

inbreeding and outbreeding. In D. B. Goldstein, and C. Schlotterer,

eds. Microsatellites: Evolution and Applications, pp. 151–164.

Oxford University Press, Oxford.

Primmer, C. R., and J. Merila. 2002. A low rate of cross-species microsat-

ellite amplification success in Ranid frogs. Conservation Genetics

3:445–449.

Raymond, M., and F. Rousset. 1995. genepop version 1.2: population

genetics software for exact tests and ecumenicism. Journal of


Richards-Zawacki, C. L. 2009. Effects of slope and riparian habitat con-

nectivity on gene flow in an endangered Panamanian frog, Atelopus

varius. Diversity and Distributions 15:796–806.

Rowe, G., and T. J. C. Beebee. 2003. Population on the verge of a

mutational meldtown? Fitness cost of genetic load for an amphibian

in the wild. Evolution 57:177–181.

Saccheri, I., and I. Hanski. 2006. Natural selection and population

dynamics. Trends in Ecology & Evolution 21:341–347.

Saccheri, I., M. Kuussaari, M. Kankare, P. Vikman, W. Fortelius, and

I. Hanski. 1998. Inbreeding and extinction in a butterfly metapopu-

lation. Nature 392:491–494.

Salvidio, S. 2009. Detecting amphibian population cycles: the impor-

tance of appropriate statistical analyses. Biological Conservation

142:455–461.

Santucci, F., K. M. Ibrahim, A. Bruzzone, and G. M. Hewit. 2007.

Selection on MHC-linked microsatellite loci in sheep populations.


Van Oosterhout, C., M. K. Van Heuven, and P. M. Brakefield. 2004.

On the neutrality of melecular genetic markers:pedigree analysis of

genetic variation in fragmented populations. Molecular Ecology

13:1025–1034.

Vieites, D. R., S. Nieto-Roman, M. Barluenga, A. Palanca, M. Vences,

and A. Meyer. 2004. Post-mating clutch piracy in an amphibian.

Nature 431:305–308.

Vonesh, J. R., and O. De la Cruz. 2002. Complex life cycles and den-

sity dependance: assessing the contribution of egg mortality to

amphibian declines. Oecologia 133:325–333.

Wang, J. 2000. Effects of population structures and selection strategies

on the purging of inbreeding depression due to deleterious muta-

tions. Genetical Research 76:75–86.

Wang, J., W. G. Hill, D. Charlesworth, and B. Charlesworth. 1999.

Dynamic of inbreeding depression due to deleterious mutations in

small populations: mutation parameters and inbreeding rate. Geneti-

cal Research 74:165–178.

Wright, S. 1969. Evolution and the Genetic of Populations. Vol. 2: The

Theory of Gene Frequencies. University of Chicago Press, Chicago.

Wright, L. I., T. Tregenza, and D. J. Hosken. 2008. Inbreeding, inbreed-

ing depression and extinction. Conservation Genetics 9:833–843.

Zartman, C., S. F. McDaniel, and A. J. Shaw. 2006. Experimental

habitat fragmentation increases linkage disequilibrium but does not

affect genetic diversity or population structure in the Amazonian

liverworth Radula flaccida. Molecular Ecology 15:2305–2315.



rapid selection against inbreeding in a wild population of a rare frog

Documents